The apical ectodermal ridge (AER), a rim of thickened ectodermal cells at the interface between the dorsal and ventral domains of the limb bud, is required for limb outgrowth and patterning. We have previously shown that the limbs of En1 mutant mice display dorsal-ventral and proximal-distal abnormalities, the latter being reflected in the appearance of a broadened AER and formation of ectopic ventral digits. A detailed genetic analysis of wild-type, En1 and Wnt7a mutant limb buds during AER development has delineated a role for En1 in normal AER formation. Our studies support previous suggestions that AER maturation involves the compression of an early broad ventral domain of limb ectoderm into a narrow rim at the tip and further show that En1 plays a critical role in the compaction phase. Loss of En1 leads to a delay in the distal shift and stratification of cells in the ventral half of the AER. At later stages, this often leads to development of a secondary ventral AER, which can promote formation of an ectopic digit. The second AER forms at the juxtaposition of the ventral border of the broadened mutant AER and the distal border of an ectopic Lmx1b expression domain. Analysis of En1/Wnt7a double mutants demonstrates that the dorsalizing gene Wnt7a is required for the formation of the ectopic AERs in En1 mutants and for ectopic expression of Lmx1b in the ventral mesenchyme. We suggest a model whereby, in En1 mutants, ectopic ventral Wnt7a and/or Lmx1b expression leads to the transformation of ventral cells in the broadened AER to a more dorsal phenotype. This leads to induction of a second zone of compaction ventrally, which in some cases goes on to form an autonomous secondary AER.

Classical tissue transplantation experiments and recent genetic studies in the limb indicate that ‘signalling centers’ provide the molecular cues necessary for patterning the limb and directing regional outgrowth along the three different axes (Saunders, 1948 and/or Martin, 1995; Johnson et al., 1994; Tickle, 1995; Niswander, 1996; Zeller and Duboule, 1997). One such signalling center is the apical ectodermal ridge (AER), a narrow rim of thickened ectoderm along the limb margin. The AER is critical for normal proximal-distal limb outgrowth (Saunders, 1948) and this function appears to be mediated by members of the fibroblast growth factor family, such as Fgf4 and Fgf8 (Niswander et al., 1993; Ohuchi et al., 1995; Mahmood et al., 1995; Crossley et al., 1996; Vogel et al., 1996). Patterning along the anterior-posterior (A-P) axis is specified by the zone of polarizing activity (ZPA), a cluster of mesenchymal cells in the posterior limb bud (Saunders and Gasseling, 1968). A major component of the ZPA signalling cascade is the protein Sonic hedgehog (Shh) (Riddle et al., 1993). Dorsal-ventral (D-V) axis polarity is influenced by the non-AER ectoderm (Pautou and Kieny, 1973; MacCabe et al., 1974; Pautau, 1977; Geduspan and MacCabe, 1987, 1989; Akita, 1996). Wnt7a, a secreted factor expressed by the dorsal limb ectoderm, is required for the differentiation of dorsal mesenchymal structures (Parr et al., 1993; Parr and McMahon, 1995) and mis-expression studies in chick suggest it acts by inducing expression of the LIM homeodomain-containing gene, Lmx1b, in the underlying mesenchyme (Riddle et al., 1995; Vogel et al., 1995). Engrailed-1 (En1), a homeodomain-containing transcriptional regulator expressed in the ventral limb ectoderm (Davis et al., 1991; Gardner and Barald, 1992), is critical for directing the development of ventral limb structures and appears to act in part by suppressing Wnt7a function in the ventral limb (Loomis et al., 1996; Logan et al., 1997). Recent studies have shown that the function of these different signalling regions is interdependent. Shh signalling by the ZPA is required for maintenance of the AER (Laufer et al., 1994; Niswander et al., 1994), and Fgf4 and Wnt7a expression are required for maintenance of Shh expression ((Laufer et al., 1994; Niswander et al., 1994; Parr and McMahon, 1995; Yang and Niswander, 1995).

Our studies of mice homozygous for En1 loss-of-function alleles suggested that an additional regulatory relationship exists between D-V limb patterning genes and AER formation. En1 mutant limbs not only demonstrate a partial dorsal transformation of the ventral paw, but they also display a ventroproximal expansion of the AER (Loomis et al., 1996). We suggested that this expansion of the newly formed AER could be responsible for the presence of ectopic ventral digits on the proximal paws of some mutants. In keeping with our findings, recent misexpression studies in chick have shown that ectopic expression of En1 in the AER and dorsal chick limb either ablates normal AER development or results in ectopic secondary AERs (Laufer et al., 1997; Logan et al., 1997; Rodriguez-Esteban et al., 1997). Recent mutant and mis-expression experiments in chick (Laufer et al., 1997; Rodriguez-Esteban et al., 1997), as well as mutant studies in mouse (Sidow et al., 1997), indicate that Notch signaling also plays a role in AER development. One of the molecules in this pathway, R-fng, has been shown to be dorsally restricted in chick limbs and appears to be regulated in part by En1 (Laufer et al., 1997; Rodriguez-Esteban et al., 1997).

Correlative molecular data in several chick mutants also suggest a link between D-V patterning and AER development. In the chick limbless mutant in which limb outgrowth does not occur due to failure to form an AER (Prahlad et al., 1979), molecular studies demonstrate that expression of AER marker genes is not initiated and normal D-V axis asymmetries are not established (Grieshammer et al., 1996; Noramly et al., 1996; Ros et al., 1996). Expression of Wnt7a and Lmx1b is expanded into the ventral limb, whereas expression of the ventralizing gene, En1, is absent in these mutants. On the contrary, chick wingless mutant limb buds initiate Fgf8 expression and display proper D-V patterning early on, but they subsequently fail to form a differentiated AER and molecular D-V asymmetries are lost (Ohuchi et al., 1997).

Based on experimental studies of Drosophila limb development and theoretical considerations in mouse, Meinhardt suggested that the juxtaposition of dorsal and ventral limb domains might be a precondition for normal vertebrate AER development (Meinhardt, 1983). In addition to the studies outlined above, further support for Meinhardt’s hypothesis is provided by recent grafting experiments in avian limbs. Chimera studies in which presumptive dorsal or ventral quail limb tissue was inserted into ventral or dorsal limb domains, respectively, resulted in formation of ectopic AERs along the borders between chick and quail tissues (Tanaka et al., 1997). Moreover, the replacement of dorsal chick ectoderm with ventral chick ectoderm at slightly later stages of limb bud development also led to the formation of ectopic AERs (Laufer et al., 1997).

In an effort to better understand the potential relationship between D-V patterning and AER formation, we have used morphological criteria as well as expression analysis of several genetic markers to compare early AER development and subsequent maturation in wild type as well as En1, Wnt7a and En1/Wnt7a double-mutant mice. Our results indicate that early mesenchymal D-V patterning is normal in En1 mutant mice whereas ventral ectoderm specification in En1 mutants is abnormal even at early stages of limb development. At later stages, mesenchymal patterning is abnormal as Lmx1b is reexpressed in the ventral mesenchyme. Moreover, maturation of the AER is aberrant resulting in an expanded ventral AER. With further development, a bifurcated AER is formed following stratification of the ventral border of the mutant AER. Interestingly, ectopic ectodermal stratification occurs at the distal edge of the ventral Lmx1b expression domain. In some cases, the (anatomically) ventral AER rim goes on to form a secondary AER capable of promoting ectopic digit development. Finally, analysis of Wnt7a/En1 double mutants demonstrates that Wnt7a is required for both the ectopic ventral expression of Lmx1b and the development of ectopic secondary AERs and ventral digits in En1 mutant limbs.

Genotype analysis of En1, Wnt7a and En1/Wnt7a mutant mice

The En1 and Wnt7a mutants used in these studies were previously generated by targeted disruption of the respective genes (Wurst et al., 1994; Hanks et al., 1995; Parr and McMahon, 1995; Matise and Joyner, 1997). In the case of the En1lki allele, the bacterial lacZ coding sequences were targeted to the En1 locus, bringing them under the control of the En1 regulatory sequences and inactivating En1 protein function (Hanks et al., 1995; Matise and Joyner, 1997). En1/Wnt7a double mutants were generated by crossing mice heterozygous for both alleles since homozygosity results in perinatal lethality and/or sterility. DNA was isolated from either yolk sacs or tails and genotyping was performed by either Southern blot analysis as previously described (Hanks et al., 1995) or PCR analysis. Primers used for genotyping the En1 wild-type allele were 5′-GTTCCAGG-CAAACCGCTATATC-3′ and for the En1hd allele were 5′-GCCAGCTCATTCCTCCCACTC-3′ and 5′-CGGTCGTAAGCAG-TTTGGCATG-3′. Primers used to genotype the Wnt7a wild-type allele were 5′-CTCTTCGGTGGTAGCTCTGG-3′ and 5′-CCTTCCC-GAAGACAGTACGC-3′ and for the Wnt7a null allele were 5′-TCACGTCCTGCACGACGCGAGCTG-3′ and 5′-CCTTCCCGA-AGACAGTACGC-3′ (Cygan et al, 1997).

Histology and in situ hybridizations

Digoxigenin-labelled whole-mount and 35S-labelled section RNA in situ hybridizations were performed essentially as described (Loomis et al., 1996; Matise and Joyner, 1997). The probes used were to En1 (Wurst et al., 1994), Wnt7a (Parr and McMahon, 1995), Fgf8 (Crossley and Martin, 1995), Fgf4 (Niswander and Martin, 1992), Dlx2 (Robinson and Mahon, 1994), Jag1 (Lindsell et al., 1995) and Lmx1b (Cygan et al., 1997). Lmx1b is the mouse homolog of the previously characterized chick Lmx1 gene (Cygan et al., 1997). Embryos for β-galactosidase staining were fixed for 20 minutes to 1 hour, depending on the age, in 4% paraformaldehyde in phosphate-buffered saline on ice and then stained with either X-gal (Sigma) or salmon-gal (Biosynth) substrates diluted in a standard staining buffer containing both 0.02% NP-40 and 0.01% deoxycholate and incubated overnight at room temperature. Some salmon-gal stained embryos were used for double labelling by RNA in situ hybridization. These embryos were postfixed for 24 hours in 4% paraformaldehyde at 4°C and then hybridized with RNA probes as above, omitting the bleaching and methanol dehydration steps (Matise and Joyner, 1997). Proteinase K treatments were also reduced so that embryos were incubated in either 2 μg/ml or 0.5 μg/ml for 5 minutes at 37°C for probes to mesenchymal or ectodermal mRNAs, respectively. Cryosections of embryos used in whole-mount RNA in situ hybridizations were prepared by first sinking the embryos in 15% sucrose overnight followed by 30% sucrose for 4 hours. The embryos were then embedded in OCT medium and cut at 10-40 μm.

D-V and AER marker gene expression in early pre-AER limb buds

Our previous analysis of the En1 mutant AER defect focused on the limb bud stage when a definitive, multilayered AER arises in wild-type mice (10.5 dpc for the forelimb and 11.5 dpc for the hindlimb). To begin to understand the genetic events that trigger and sustain the development of an aberrant AER in En1 mutants, we performed a detailed analysis of the expression patterns of AER and D-V marker genes in wild-type and mutant limb buds. For consistency, we adopted the mouse limb staging system first developed by Wanek et al. (1989) and modified by Bell and Scott (personal communication). At stage 0.25, the limb bud is not yet morphologically discernable along the embryo flank but, by stage 0.5, it is delineated by a slight prominence and its ventral ectoderm is slightly thickened compared to the flattened dorsal ectoderm. At stage 1, the limb is a broad bulge with the distoventral ectoderm thickened relative to the dorsal and more proximal ventral ectoderm (see Fig. 1B,E,H,K). At stage 2, the limb bud looks like a semicircular protrusion emanating from the flank, and a broad crescent or band of bilayered ectoderm is present on the distal third of the ventral limb (Milaire, 1974). At stage 3 (10.5 dpc in the forelimb), the definitive, mature AER is visible along the distal rim of the elongated, cylindrical limb bud and consists of 3-4 layers of ectodermal cells (see Fig. 1C,F,I,L). In mouse, the AER appears to form a true stratified epithelium (Kelley and Fallon, 1983).

Fig. 1.

The position of D-V and AER marker gene expression shifts dorsodistally during the transition from a pre-AER to an AER stage limb bud. Whole-mount RNA in situ hybridization and subsequent sagittal cryosections of wild-type limbs at stage 1 (A,B,D,E,G,H,J,K) and stage 3 (C,F,I,L). Dorsal (d) and ventral (v) as indicated. (A-C) En1 expression in the early limb bud. Note the distal shift of En1 expression from a proximoventral position at stage 1 (A) to a broad region encompassing both the ventral ectoderm and ventral AER at stage 3 (B,C). (D,F) Fgf8 and (E) Dlx2 expression was detected in a broad domain of thickened ectoderm along the distal half of the ventral limb bud at stage 1 (D,E), but by stage 3 this expression domain had both compressed and shifted distally, delineating the definitive AER (F). Note that the position of morphologically thickened ectoderm shifted in parallel with AER marker gene expression. (G-I) Early Wnt7a expression was detected throughout much of the dorsal and ventral limb ectoderm but, by stage 3, had become restricted to the dorsal surface. Lmx1b expression (J-L) was initially in a broad region that included ventral, distal and dorsal mesenchyme but, by stage 3, had became restricted to the dorsal mesenchyme. Mesenchymal staining is limited to a superficial band due to incomplete penetration of the probe. Bar: 100 μm.

Fig. 1.

The position of D-V and AER marker gene expression shifts dorsodistally during the transition from a pre-AER to an AER stage limb bud. Whole-mount RNA in situ hybridization and subsequent sagittal cryosections of wild-type limbs at stage 1 (A,B,D,E,G,H,J,K) and stage 3 (C,F,I,L). Dorsal (d) and ventral (v) as indicated. (A-C) En1 expression in the early limb bud. Note the distal shift of En1 expression from a proximoventral position at stage 1 (A) to a broad region encompassing both the ventral ectoderm and ventral AER at stage 3 (B,C). (D,F) Fgf8 and (E) Dlx2 expression was detected in a broad domain of thickened ectoderm along the distal half of the ventral limb bud at stage 1 (D,E), but by stage 3 this expression domain had both compressed and shifted distally, delineating the definitive AER (F). Note that the position of morphologically thickened ectoderm shifted in parallel with AER marker gene expression. (G-I) Early Wnt7a expression was detected throughout much of the dorsal and ventral limb ectoderm but, by stage 3, had become restricted to the dorsal surface. Lmx1b expression (J-L) was initially in a broad region that included ventral, distal and dorsal mesenchyme but, by stage 3, had became restricted to the dorsal mesenchyme. Mesenchymal staining is limited to a superficial band due to incomplete penetration of the probe. Bar: 100 μm.

Molecular analysis of wild-type limb buds prior to the formation of a morphologically distinct AER demonstrated dynamic shifts in the topological location of both D-V and AER marker gene expression domains. Between stages 0.25 and 0.5, the AER marker gene Dlx2 was expressed throughout the ventral ectoderm (data not shown) but, by stage 1, the Fgf8/Dlx2 expression domain occupied only the distal half of the ventral limb bud (Fig. 1D,E). Over the next few stages, this domain was compressed even further so that Dlx2 and Fgf8 expression was ultimately restricted to the multilayered AER along the D-V interface (Fig. 1F). Over the same stages, the En1 expression domain shifted distally, from a relatively restricted proximoventral position to extend over the entire ventral limb including the ventral AER (Fig. 1A-C). In parallel with these relative distal shifts in the topological location of AER and ventral ectoderm marker genes, the ‘dorsal’ marker genes, Wnt7a and Lmx1, were also displaced more dorsally (Fig. 1G-L). Wnt7a, which at stage 1 was expressed in the distal and dorsal ectoderm covering 2/3 of the limb, became restricted to only the dorsal half of the limb ectoderm by stage 3 (Fig. 1H,I). Lmx1b, which was found to be expressed in both the ventral and dorsal limb mesenchyme prior to stage 2, became restricted to the dorsal mesenchyme by stage 3 (Fig. 1K,L).

Differentiation and compression of the ventral AER are aberrant in En1 mutant embryos

To determine if En1 plays a role in setting up the expression domains in the early limb bud, we conducted the same analysis in En1 mutant embryos. D-V ectodermal patterning was abnormal even at the earliest limb stages. Ectopic ventral expression of Wnt7a could be detected as early as stage 0.5 in En1 mutants (data not shown), with intense expression persisting in the ventroproximal ectoderm but clearing from the ventrodistal ectoderm of older pre-AER and AER stage limbs (Fig. 2 compare A and B, E and F). In contrast, expression of the mesenchymal gene Lmx1b was normal in En1 mutant limbs at these early stages as it was appropriately expressed in early dorsal and ventral mesenchyme and then repressed despite persistent expression of Wnt7a in the adjacent ventral ectoderm (Fig. 2I,J and data not shown).

Fig. 2.

Ectodermal D-V patterning and early AER differentiation is abnormal in En1 mutant limbs. (A-J) Whole-mount and (K,L) section RNA in situ hybridizations of limbs from wild-type or En1 heterozygous mice (A,C,E,G,I,K) and En1 mutant mice (B,D,F,H,J,L) at stages 1/2 (A-D) and 3/4 (E-L). Dorsal (d) and ventral (v) as indicated. (A,B,E,F) Wnt7a probe illustrating the ectopic ventral expression in mutant limb buds (B,F) compared to wild-type limbs (A,E). Note the central clearing of Wnt7a expression from the ventrodistal ectoderm in the mutant limb buds (B,F). (C,D,G) Dlx2 and (H) Fgf8 probes demonstrating the marked ventroproximal expansion of the expression domain in mutant embryos (D,H), compared to wild-type embryos (C,G). Note that at stage 3, the AER, as delineated by Dlx2 expression, is slightly broader anteriorly (arrow) than posteriorly (arrowhead). (I,J) Double-label whole mounts for En1-lacZ expression (pink) and Lmx1b expression (purple) demonstrating that cells expressing En1-lacZ are present throughout the ventral limb ectoderm, extending to the distal tip, and that Lmx1b expression is dorsally restricted in stage 3 limbs from En1 heterozygotes (I) and mutants (J). Note the clear distal area devoid of probe which demarcates the location of the dorsal AER. The greater intensity of β-galactosidase staining in mutant compared to heterozygous limbs is due in part to the presence of two lacZ-targeted En1 alleles. Section in situ hybridization of stage 3/4 (K) wild-type and (L) En1 mutant limbs using a 35S-labelled Fgf4 probe. Note the restriction of Fgf4 expression to the dorsal-most portion of the partially stratified mutant AER (L) compared to its expression throughout the dorsal and ventral halves of the wild-type AER (K). (M) Schematic illustrating the expression pattern of these genes in stage 1/2 and stage 3/4 limb buds. Bar: 100 μm.

Fig. 2.

Ectodermal D-V patterning and early AER differentiation is abnormal in En1 mutant limbs. (A-J) Whole-mount and (K,L) section RNA in situ hybridizations of limbs from wild-type or En1 heterozygous mice (A,C,E,G,I,K) and En1 mutant mice (B,D,F,H,J,L) at stages 1/2 (A-D) and 3/4 (E-L). Dorsal (d) and ventral (v) as indicated. (A,B,E,F) Wnt7a probe illustrating the ectopic ventral expression in mutant limb buds (B,F) compared to wild-type limbs (A,E). Note the central clearing of Wnt7a expression from the ventrodistal ectoderm in the mutant limb buds (B,F). (C,D,G) Dlx2 and (H) Fgf8 probes demonstrating the marked ventroproximal expansion of the expression domain in mutant embryos (D,H), compared to wild-type embryos (C,G). Note that at stage 3, the AER, as delineated by Dlx2 expression, is slightly broader anteriorly (arrow) than posteriorly (arrowhead). (I,J) Double-label whole mounts for En1-lacZ expression (pink) and Lmx1b expression (purple) demonstrating that cells expressing En1-lacZ are present throughout the ventral limb ectoderm, extending to the distal tip, and that Lmx1b expression is dorsally restricted in stage 3 limbs from En1 heterozygotes (I) and mutants (J). Note the clear distal area devoid of probe which demarcates the location of the dorsal AER. The greater intensity of β-galactosidase staining in mutant compared to heterozygous limbs is due in part to the presence of two lacZ-targeted En1 alleles. Section in situ hybridization of stage 3/4 (K) wild-type and (L) En1 mutant limbs using a 35S-labelled Fgf4 probe. Note the restriction of Fgf4 expression to the dorsal-most portion of the partially stratified mutant AER (L) compared to its expression throughout the dorsal and ventral halves of the wild-type AER (K). (M) Schematic illustrating the expression pattern of these genes in stage 1/2 and stage 3/4 limb buds. Bar: 100 μm.

AER marker gene expression was initiated normally in En1 mutant limb buds, but failed to shift distally as occurs in wild-type limb buds. The Fgf8/Dlx2 expression domain in stage 1 En1 mutant limbs was clearly broader than in wild-type limbs (Fig. 2, compare C and D). By stage 3/4, although the proximal boundary of the Dlx2/Fgf8-positive domain had shifted distally slightly, the mutant AER remained much broader than its wild-type counterpart (Fig 2, compare G and H). Expression of Fgf4, a late stage AER differentiation marker, was mainly restricted to the dorsal-most cells of the En1 mutant AER (Fig. 2K,L).

To determine whether the expanded portion of the En1 mutant AER corresponded primarily to the ventral AER, which normally expresses En1, we analyzed mice heterozygous or homozygous for an En1 targeted null allele, En1lki (Hanks et al., 1995; Matise and Joyner, 1997), which expresses the lacZ reporter gene from the En1 locus. Double labelling for β-galactosidase activity (En1-lacZ expression) and Lmx1b mRNA showed that, by stage 3 in both En1lki/+ and En1lki/lki mutants, the size of the dorsal AER, delineated as the ‘negative’ domain between Lmx1b expression dorsally and lacZ ventrally, was very similar (Fig. 2I,J). These observations, together with the Fgf4 expression data and our previous molecular and morphological studies (Loomis et al., 1996), indicate that the dorsal AER is not grossly altered in En1 homozygous mutants; En1 appears to be required primarily for compaction and maturation of the ventral AER.

Development of a second ectodermal thickening along the proximoventral boundary of the expanded En1 mutant AER

Since the ventroproximally expanded AER detected in En1 mutant forelimb buds at stage 3 (Loomis et al., 1996) resembled the thickened ectoderm normally seen in a stage 2 wild-type limb, one question was whether the ventral expansion of the En1 mutant AER at 10.5 dpc simply reflected a delay in the distal shift of cells expressing Dlx2/Fgf8. To address this, we compared AER morphology and marker gene expression patterns in wild-type and mutant limbs at later stages of maturation and found that, over the next several stages, the En1 mutant AER expression domain remained broader than the wild-type AER expression domain (compare Fig. 3A,C with B,D). Surprisingly, whole-mount gene expression analyses revealed that, at limb stages 5/6, the ventral expansion of the En1 mutant AERs exhibited a marked A-P asymmetry (Fig. 3B), which was not prominent at limb stages 3/4 (Fig 2D). The anterior half of the En1 mutant AER at these stages was up to 4-6 times wider than the normal AER, whereas the posterior half was usually only 2-3 times wider than the wild type. This A-P asymmetry was observed in 90% of limbs analyzed (n>40) using a variety of AER-specific probes.

Fig. 3.

Loss of En1 function results in reexpression of Lmx1b ventrally and abnormal AER maturation, leading in some cases to formation of ectopic AERs. Whole-mount in situ hybridization and/or β-galactosidase staining of stage 5/6 (A-H,J) and stage 7/8 (I,K,L) wild-type (A,C,E,G), En1 mutant (B,D,F,H,J-L) and Wnt7a mutant limbs (I). Dorsal (d) and ventral (v) are indicated. Dlx2 probe (A,B) illustrating the intense staining of the dorsal and ventral rims of the bifurcated mutant AER and the marked widening anteriorly (arrow, B) relative to posteriorly (arrowhead, B) and compared to the narrow ridge of a wild-type AER (A). β-galactosidase staining (C,D) demonstrating the strong expression of En1-lacZ in the AER and weaker staining in the ventral ectoderm of limbs from (C) heterozygous and (D) homozygous mice. Again note the mutant AER is wider anteriorly (arrow) versus posteriorly (arrowhead). Lmx1b probe (E,F) illustrating the normal dorsal expression (E) and the ectopic reexpression in the ventral mesenchyme of En1 mutant limbs (F). Note that the contours of the distal Lmx1b-negative zone have the same contours of an expanded AER. Jagged-1 expression (G,H), which was restricted to the progress zone of wild-type limb buds (G), extended into these ventral mesenchymal bulges associated with the marked anterior widening of the En1 mutant AERs (H, arrow). Lmx1b expression (purple) in a stage 5/6 Wnt7a mutant hindlimb (I, arrowhead) is relatively normal and nearly abuts the En1-lacZ marked AER (pink). However, a distal Lmx1b-negative zone is clearly present in the more mature, stage 7/8 forelimb (I, arrow). Fgf4 expression (J) in the dorsal rim and a small region of the ventral rim of the bifurcated En1 mutant AER (arrow). Fgf8 expression (purple,K,L) reveals persistent focal expression (arrows) in the ventral ectoderm of some En1 mutant limbs at about the time when Fgf8 transcripts are down-regulated in the primary AER, which is delineated by intense En1-lacZ expression (pink). These secondary AER signalling centers always overlie ectopic mesenchymal outgrowths. Bar: 500 μm.

Fig. 3.

Loss of En1 function results in reexpression of Lmx1b ventrally and abnormal AER maturation, leading in some cases to formation of ectopic AERs. Whole-mount in situ hybridization and/or β-galactosidase staining of stage 5/6 (A-H,J) and stage 7/8 (I,K,L) wild-type (A,C,E,G), En1 mutant (B,D,F,H,J-L) and Wnt7a mutant limbs (I). Dorsal (d) and ventral (v) are indicated. Dlx2 probe (A,B) illustrating the intense staining of the dorsal and ventral rims of the bifurcated mutant AER and the marked widening anteriorly (arrow, B) relative to posteriorly (arrowhead, B) and compared to the narrow ridge of a wild-type AER (A). β-galactosidase staining (C,D) demonstrating the strong expression of En1-lacZ in the AER and weaker staining in the ventral ectoderm of limbs from (C) heterozygous and (D) homozygous mice. Again note the mutant AER is wider anteriorly (arrow) versus posteriorly (arrowhead). Lmx1b probe (E,F) illustrating the normal dorsal expression (E) and the ectopic reexpression in the ventral mesenchyme of En1 mutant limbs (F). Note that the contours of the distal Lmx1b-negative zone have the same contours of an expanded AER. Jagged-1 expression (G,H), which was restricted to the progress zone of wild-type limb buds (G), extended into these ventral mesenchymal bulges associated with the marked anterior widening of the En1 mutant AERs (H, arrow). Lmx1b expression (purple) in a stage 5/6 Wnt7a mutant hindlimb (I, arrowhead) is relatively normal and nearly abuts the En1-lacZ marked AER (pink). However, a distal Lmx1b-negative zone is clearly present in the more mature, stage 7/8 forelimb (I, arrow). Fgf4 expression (J) in the dorsal rim and a small region of the ventral rim of the bifurcated En1 mutant AER (arrow). Fgf8 expression (purple,K,L) reveals persistent focal expression (arrows) in the ventral ectoderm of some En1 mutant limbs at about the time when Fgf8 transcripts are down-regulated in the primary AER, which is delineated by intense En1-lacZ expression (pink). These secondary AER signalling centers always overlie ectopic mesenchymal outgrowths. Bar: 500 μm.

In addition to its marked anterior expansion, the En1 mutant AER also appeared to bifurcate into two distinct ectodermal ridges between stages 5 and7. In situ staining with Fgf8 or Dlx2 riboprobes outlined two intensely stained rims of ectoderm, which were separated by an intervening region of less intensely stained epithelium (Fig. 3B). The dorsal rim was anchored at the distal limb margin whereas the ventral rim formed along the ventroproximal boundary of the expanded Dlx2/Fgf8 expression domain.

Roughly coincident with the formation of the second ventral AER rim, Lmx1b was re-expressed in the ventral mesenchyme of stage 5/6 En1 mutant limbs (Fig. 3E,F). We were somewhat surprised that ventral Lmx1b reexpression was initiated at such a late developmental stage in the ventral mesenchyme given the persistent expression of Wnt7a in the ventral ectoderm and the reported role of Wnt7a in promoting Lmx1b expression in chick (Riddle et al., 1995; Vogel et al., 1995). To more directly assess the regulatory relationship of these two genes in mouse, we analyzed Lmx1b expression in Wnt7a mutant mice. We found that Lmx1b expression was initially normal but, beginning at stages 5/6, expression was markedly decreased in the distodorsal mesenchyme adjacent to the ridge (Fig. 3I). Thus, in mouse, Wnt7a appears to regulate Lmx1b expression only at later stages of development.

Despite the intense reexpression of Lmx1b in the ventroproximal mesenchyme of En1 mutant mice, the mesenchyme directly beneath the mutant AER did not reexpress Lmx1b. The striking A-P asymmetry of the AER was paralleled by a marked A-P asymmetry in the breadth of the Lmx1b-negative mesenchymal zone. Thus, the ‘footprint’ of the widened mutant AER overlaid the Lmx1b-negative zone and appeared to be outlined by the dorsal and ventral Lmx1b expression boundaries.

Development of autonomous secondary AERs in En1 mutant limbs

The AER is a transitory embryonic structure and, by stages 6-8, most AER-specific genes, including Fgf4 and Fgf8, are down-regulated and the AER begins to flatten and regress (Jurand, 1965; Milaire, 1974; Wanek et al., 1989). In some stage 6-8 En1 mutant limbs, however, clusters of cells associated with the ventral AER rim continued to express AER marker genes when they were fading elsewhere (Fig. 3J-L). These ectopic signalling centers were always associated with underlying mesenchymal outgrowths. As would be predicted if the signalling centers function as AERs, such mesenchymal bulges expressed genes normally restricted to cells of the progress zone that underlie the AERs, such as Evx1 (Niswander and Martin, 1993) and Jagged-1 (Fig. 3G,H and data not shown). Shh, a component of the pathway directing A-P patterning of the distal limb, was rarely detected in the ectopic outgrowths (data not shown). Although almost all En1 mutant limbs demonstrated a marked ventral expansion of the anterior half of the bifurcated AER, only 30-40% (n>40) of mutant limbs gave rise to secondary signalling centers with persistent AER-marker expression. In addition, the formation of such signalling domains occurred only in anterior regions near the widest separation between the dorsal and ventral AER rims.

By limb stages 9-11, the normal regressing AER is difficult to distinguish morphologically from the adjacent ventral and dorsal ectoderm (Wanek et al., 1989). Interestingly, analysis of β-galactosidase activity in limbs of mice carrying the En1lki allele, suggested that a remnant ridge persists along the D-V boundary (Fig. 4A,B) up to at least stage 13 (16.5 dpc) in normal embryos. Relatively intense En1 mRNA expression also delineated a D-V boundary zone up to at least stage 12 (15.5 dpc, data not shown), indicating that the persistent β-galactosidase activity did not solely reflect stability of the β-galactosidase protein.

Fig. 4.

Further constriction of the bifurcated ridge and reorientation of the ectopic ventral AERs. Ventral views of β-galactosidase-stained heterozygous (En1lki/+) (A) and En1 mutant (En1lki/hd) (B-H) limbs between stages 9 and 12. En-lacZ expression outlining the narrow, phenotypically wild-type AER (A) at these later stages and the apparent distal shift of the ventral rim of the bifurcated En1 mutant AER (B,C). In some En1 mutant limbs at this stage, the ventral rim has become tightly apposed to the dorsal rim (B). Nevertheless, the En1 mutant AER in B stains more intensely than the phenotypically wild-type AER (A) even though both the mutant (En1lki/hd) and the (En1lki/+) heterozygotes carry a single En1lki reporter allele. In other En1 mutant limbs at this stage, the ventral rims are not yet tightly apposed to the dorsal rims, and at positions of focal mesenchymal outgrowths the ventral rims display a proximoventral bow or deviation in their otherwise linear course (D, arrowhead). With time all such bows are replaced by secondary linear ridges which join the AER at a 90° angle (E,F). Secondary ridges which developed more proximally overlie differentiating ectopic digits and appear to be discontinuous with the primary ridge at this stage (G, arrowhead). However, if viewed from the side, some are found to be linked to the primary AER through a residual mesenchymal thickening (H, arrow). Lmx1b expression (purple) illustrates the double dorsal nature of a ventral outgrowth in a stage 10 limb bud (I). En1-lacZ expression (pink) demarcates the position of the ectopic secondary AER. The two dorsal domains are on the anterior (a) and posterior aspects of the ectopic outgrowth and the proximal-distal axis is perpendicular to the primary P-D axis. Note the absence of Lmx1b staining in the interdigital region (arrowhead, I). By these later time points, Lmx1b expression has cleared (arrowhead, J) from the posterior (p) ventral hand plate. En-lacZ (blue) expression illustrating an early ventral digit in a stage 12 En1 mutant limb (K). Hematoxylin and eosin stained sagittal sections of a well differentiated En1 mutant ventral digit (L) demonstrating nail plates (n) on both the dorsal and ventral surfaces, confirming the double-dorsal nature of these structures. Bar: 500 μm.

Fig. 4.

Further constriction of the bifurcated ridge and reorientation of the ectopic ventral AERs. Ventral views of β-galactosidase-stained heterozygous (En1lki/+) (A) and En1 mutant (En1lki/hd) (B-H) limbs between stages 9 and 12. En-lacZ expression outlining the narrow, phenotypically wild-type AER (A) at these later stages and the apparent distal shift of the ventral rim of the bifurcated En1 mutant AER (B,C). In some En1 mutant limbs at this stage, the ventral rim has become tightly apposed to the dorsal rim (B). Nevertheless, the En1 mutant AER in B stains more intensely than the phenotypically wild-type AER (A) even though both the mutant (En1lki/hd) and the (En1lki/+) heterozygotes carry a single En1lki reporter allele. In other En1 mutant limbs at this stage, the ventral rims are not yet tightly apposed to the dorsal rims, and at positions of focal mesenchymal outgrowths the ventral rims display a proximoventral bow or deviation in their otherwise linear course (D, arrowhead). With time all such bows are replaced by secondary linear ridges which join the AER at a 90° angle (E,F). Secondary ridges which developed more proximally overlie differentiating ectopic digits and appear to be discontinuous with the primary ridge at this stage (G, arrowhead). However, if viewed from the side, some are found to be linked to the primary AER through a residual mesenchymal thickening (H, arrow). Lmx1b expression (purple) illustrates the double dorsal nature of a ventral outgrowth in a stage 10 limb bud (I). En1-lacZ expression (pink) demarcates the position of the ectopic secondary AER. The two dorsal domains are on the anterior (a) and posterior aspects of the ectopic outgrowth and the proximal-distal axis is perpendicular to the primary P-D axis. Note the absence of Lmx1b staining in the interdigital region (arrowhead, I). By these later time points, Lmx1b expression has cleared (arrowhead, J) from the posterior (p) ventral hand plate. En-lacZ (blue) expression illustrating an early ventral digit in a stage 12 En1 mutant limb (K). Hematoxylin and eosin stained sagittal sections of a well differentiated En1 mutant ventral digit (L) demonstrating nail plates (n) on both the dorsal and ventral surfaces, confirming the double-dorsal nature of these structures. Bar: 500 μm.

β-galactosidase activity in En1lki/lki mice was used to follow the evolution of the mutant ectopic AERs. As shown in Fig. 4C, the anterior ventral rim of a stage 8/9 mutant limb was shifted more distally than what was observed at earlier developmental stages, thus positioning the ventral rim in closer proximity to the dorsal rim. In some mutant limbs at this stage (n=5/16 limbs analyzed), the dorsal and ventral rims had become very closely apposed, producing a narrow linear ridge along the D-V interface, which was only slightly wider than the wild-type counterpart (Fig. 4A,B).

In other mutant limbs (n=8/16), however, a region of the ventral ridge remained associated with proximal mesenchymal outgrowths and, as a result, formed a small deflection or bow in the otherwise linear contour of the ventral rim (Fig 4D). Over the next few developmental stages, such bows were completely replaced by linear ridges (n=3/16 at stages 8-9 and n=6/12 at stages 10-12) which intersected the primary ridge at a 90° angle (Fig. 4E,F). In some cases, a proximal segment of the ventral AER rim became discontinuous with the primary rim and was at an approximately 90° angle to the primary rim (5/30, stages 8-12). These more proximal, autonomous ridges always overlaid elongated outgrowths, which resembled developing digits (Fig 4G,H,K). Intermediate stages could occasionally be detected in which such autonomous En1-lacZ-positive ridges could be shown to be indirectly connected to the primary ridge through a ventral mesenchymal thickening (Fig. 4H).

Lmx1b expression boundaries are associated with ectopic secondary AERs

Lmx1b limb expression, like En1, persisted long after most AER-specific marker genes and Wnt7a expression had faded. By stage 9, Lmx1b expression was closely associated with the developing dorsal skeletal elements of the foot and had cleared from the interdigital webs of the dorsal domains of both wild-type and mutant limbs (Fig. 4I).

The ventral Lmx1b expression pattern in En1 mutant limb buds at these later stages demonstrated several striking features. First, the pattern of interdigital web clearing of Lmx1b was more irregular on the ventral limb than on the dorsal limb. Second, Lmx1b expression was not maintained in the posterior ventral footplate (Fig. 4J). This observation was consistent with our earlier observation that the posterior-most digits of En1 mutant limbs were less dorsalized than the more anterior digits (Loomis et al., 1996). Second, there was dramatic clearing of Lmx1b expression below the ectopic AERs (Fig. 4I). This was similar to the striking Lmx1b-negative footprint produced by the widened primary AER at earlier stages in the En1 mutant limbs. Finally, the ectopic ventral outgrowths expressed Lmx1b on both sides of the Lmx1b-negative zone beneath the ectopic AERs, suggesting they had a double-dorsal identity (Fig. 4I). Interestingly, the two Lmx1b-positive dorsal domains were on the anterior and posterior sides of the ectopic digit relative to the rest of the limb, reflecting the 90° rotation of the ectopic ridge. In rare En1 homozygotes that survive past birth, it was clear that morphologically, the well-differentiated ventral digits have cylindrical nails similar to the En1 double-dorsal digits (Fig. 4L).

Wnt7a is required for ectopic ventral digit formation in En1 mutants

Our expression studies implicated the ectopic expression of Wnt7a and Lmx1b in the formation of ectopic AERs and digit formation. To further investigate this association, En1/Wnt7a double homozygous mutants were constructed by interbreeding En1 and Wnt7a heterozygotes (Wurst et al., 1994; Hanks et al., 1995; Parr and McMahon, 1995; Matise and Joyner, 1997). In En1/Wnt7a double mutants, as in the Wnt7a mutants, the AERs were found to be slightly (1.5-2 fold) wider than wild-type AERs (Fig. 5A,C,D), and staining for En1-lacZ expression suggested that the dorsal, rather than the ventral, portion of the AER was primarily expanded in these mutants (data not shown). None of the stage 3/4 En1/Wnt7a double mutant AERs (n=12) displayed the dramatic ventral expansion characteristic of En1 mutant AERs (Fig. 5B). Also, none of the AERs of stage 5-8 En1/Wnt7a double mutants developed a second ventral rim or appeared bifurcated (n=16). This is in contrast to the >90% of En1 mutant limbs which display these features. Finally, none of the stage 7-10 En1/Wnt7a double-mutant embryonic limbs displayed ventral mesenchymal outgrowths associated with functional ectopic AERs (n=8) nor did postnatal double mutants have ectopic ventral digits (n=30). Thus, ectopic Wnt7a signalling in the ventral ectoderm appears to be essential for the delay in compaction of the ventral En1 mutant AER and for subsequent ectopic AER and digit formation.

Fig. 5.

Formation of an En1 mutant AER and certain D-V patterning defects requires Wnt7a expression. (A-D) Whole-mount RNA in situ hybridizations of stage 4 wild-type (A), En1 mutant (B), Wnt7a mutant (C) and En1/Wnt7a double-mutant (D) limb buds probed with Fgf8 (purple). Some of the embryos were also stained for En1-lacZ (pink). Note, that the Wnt7a and En1/Wnt7a mutant AERs are slightly wider than the wild-type AER, but much narrower than the En1 mutant AER. Whole-mount dorsal view (E-G) and high power ventral view (H-J) of young adult wild-type (E,H), En1 mutant (I), Wnt7a mutant (F) and En1/Wnt7a double-mutant (G,J) paws. The dorsal surface of a wild-type paw is covered with hair, but lacks dermal pads (E). In contrast, the dorsal surface of Wnt7a and En1/Wnt7a double-mutant paws (F,G) show variable loss of hair follicles distally and develop ectopic pigmented pads, which grow hard, nail-like structures (arrowheads). In contrast to the ectopic dorsal pads, wild-type ventral pads (H) are soft, compressible bulges with a non-shiny epidermal surface which lacks pigmentation except at the opening of sweat gland ducts. The ventral pads as well as the transverse digit ridges of En1 mutant (I) and En1/Wnt7a (J) double-mutant paws, however, are firm, shiny bulges with marked pigmentation throughout the surface ectoderm. As described previously (Loomis et al., 1997) these go on to form nail-like structures in older mice similar to the ectopic dorsal pads described above. Bar: 500 μm.

Fig. 5.

Formation of an En1 mutant AER and certain D-V patterning defects requires Wnt7a expression. (A-D) Whole-mount RNA in situ hybridizations of stage 4 wild-type (A), En1 mutant (B), Wnt7a mutant (C) and En1/Wnt7a double-mutant (D) limb buds probed with Fgf8 (purple). Some of the embryos were also stained for En1-lacZ (pink). Note, that the Wnt7a and En1/Wnt7a mutant AERs are slightly wider than the wild-type AER, but much narrower than the En1 mutant AER. Whole-mount dorsal view (E-G) and high power ventral view (H-J) of young adult wild-type (E,H), En1 mutant (I), Wnt7a mutant (F) and En1/Wnt7a double-mutant (G,J) paws. The dorsal surface of a wild-type paw is covered with hair, but lacks dermal pads (E). In contrast, the dorsal surface of Wnt7a and En1/Wnt7a double-mutant paws (F,G) show variable loss of hair follicles distally and develop ectopic pigmented pads, which grow hard, nail-like structures (arrowheads). In contrast to the ectopic dorsal pads, wild-type ventral pads (H) are soft, compressible bulges with a non-shiny epidermal surface which lacks pigmentation except at the opening of sweat gland ducts. The ventral pads as well as the transverse digit ridges of En1 mutant (I) and En1/Wnt7a (J) double-mutant paws, however, are firm, shiny bulges with marked pigmentation throughout the surface ectoderm. As described previously (Loomis et al., 1997) these go on to form nail-like structures in older mice similar to the ectopic dorsal pads described above. Bar: 500 μm.

The embryonic and postnatal D-V patterning defects in the En1/Wnt7a double mutants reflect a combination of the two single mutant D-V phenotypes. As in stage 5/6 Wnt7a mutant limbs (Fig. 1), the distodorsal expression of Lmx1b is lost in En1/Wnt7a double-mutant limbs (data not shown). However, the ectopic ventral expression of Lmx1b seen in stage 5/6 En1 mutant limbs was not observed in En1/Wnt7a double mutants (data not shown), confirming the dependence of late ventral Lmx1b expression on Wnt7a signalling from adjacent ectoderm.

Postnatally, the proximal paws combined the phenotypic characteristics of both the Wnt7a and En1 single mutants, whereas the digit tips resembled those of the Wnt7a mutants. The nails of the double mutants, like those of the Wnt7a mutants, were shortened cones which were overgrown by exuberant pad tissue on both dorsal and ventral sides. As in the Wnt7a mutants, hair follicles were variably lost from the dorsal digit surfaces and foot pads, which are usually restricted to the ventral paw, developed dorsally on the En1/Wnt7a double-mutant paws (Fig. 5E-G). The ectopic dorsal pads in Wnt7a and En1/Wnt7a double mutants were distinct from wild-type ventral pads in that the dorsal pads were aberrantly pigmented, displayed a paucity of eccrine glands and developed into hard nail-like structures over time. The En1/Wnt7a double mutants also developed abnormal ventral pads similar to those observed in En1 mutant limbs (Fig. 5H-J). These pads, like the aberrant dorsal pads, were highly pigmented and developed hard nail-like surfaces over the first several weeks of life. The two compound heterozygous-homozygous limb phenotypes were indistinguishable from the respective homozygous phenotypes. This analysis demonstrates that Wnt7a is required for ectopic AER formation, as well as promoting dorsalization of the limb mesenchyme in En1 mutants. Independently of Wnt7a, En1 appears to be required for eccrine gland formation and suppression of nail-like structures in association with dermal pads during postnatal development.

Detailed analysis of D-V patterning and AER formation in wild-type and En1 and Wnt7a mutant limbs has provided further insights into the genetic and cellular pathways important in these processes. First, our results reinforce previous histological studies suggesting that mouse AER development occurs in multiple stages. Expression of early AER marker genes was found to be induced in a broad region of ectoderm covering the ventral limb and then to become restricted to the tip, and this change in the expression domain appears to coincide with compaction of the ventral ectoderm to form a morphologically visible ridge. Second, we have observed that, in mouse, unlike that reported for chick (Riddle et al., 1995; Vogel et al., 1995), the dorsalizing gene Lmx1b is initially expressed in both dorsal and ventral limb mesenchyme and then becomes dorsally restricted just prior to AER formation. In Wnt7a mutants, we observed that this early expression of Lmx1b is independent of Wnt7a. Third, our mutant analysis indicates that En1 is required during AER formation specifically for maturation of the ventral AER, where it is normally expressed. In En1 mutants, the dorsal AER appears to form normally, whereas the ventral AER does not compact resulting in an abnormally broadened AER in the ventroproximal dimension. The resulting expanded AER later bifurcates and, in some cases, forms ectopic AERs that can promote digit formation ventrally. Finally, our analysis of Wnt7a/En1 double-mutant limb buds demonstrates that Wnt7a is required for formation of ectopic AERs in En1 mutants. We suggest that this is because ectopic ventral Wnt7a and/or Lmx1b expression in En1 mutants induces ventral AER cells to take on dorsal characteristics and undergo a second compaction process ventrally.

Role of ectodermal morphogenetic movements in early AER formation

The data to date do not distinguish between whether the initial induction of the AER is coincident with formation of a morphologically thickened ridge or begins at an earlier stage. Milaire (1974) proposed that many of the important morphogenetic functions of the definitive AER are present in the ventral ectoderm of the pre-AER limb bud, based on histochemical and morphological studies of early limb buds. He suggested that mammalian AER development involves progressive thickening of the ventral ectoderm and then a convergence of these cells to the limb margin. Consistent with this, recent chick-quail chimera studies and cell-labelling experiments in chick indicate that a lineage relationship exists between cells that initially overlie a broad region of presumptive limb mesoderm and cells that later reside along the ventrodistal surface of the pre-AER limb bud and finally localize to the mature AER (Altabef et al., 1997; Michaud et al., 1997).

Our gene expression studies and morphological data provide further evidence that early morphogenetic movements are critical for normal mouse AER development (Fig. 6A). We showed that, prior to formation of a morphologically distinct AER, the AER marker genes Dlx2 and Fgf8 are expressed in a broad region of ventral ectoderm, consistent with prior observations (Lyons et al., 1990; Bulfone et al., 1993; Parr et al., 1993; Crossley and Martin, 1995; Morasso et al., 1995). We went on to further show that the ventro-proximal border of this expression domain shifts distally with increasing limb maturity and that this shift correlates temporally and spatially with the progressive morphological thickening of the ventrodistal limb ectoderm observed by Milaire (1974) (also see Fig. 1). Additionally, the En1 expression domain initially occupies only the proximal region of the early ventral limb but later extends to the distal margin. It is not clear from our studies if all the Dlx2/Fgf8-expressing cells in the ventral ectoderm of the pre-AER limb bud ultimately come to reside in the definitive AER. Indeed, recent cell-labelling studies in chick suggest that this might not be the case (Altabef et al., 1997). Nevertheless the cumulative molecular and morphological data indicate that multiple waves of ectodermal morphogenetic movements are probably critical for normal mouse AER development (Fig. 6A).

Fig. 6.

Schematic drawing illustrating a model for the morphogenetic movements important in mouse AER development. Based on our results, prior histochemical studies in mouse and lineage data in chick (Milaire,1974; Michaud et al., 1997; Altabef et al., 1997), we propose that cells which give rise to the mouse AER initially overlie much of the presumptive limb mesoderm and that sequential, convergent morphogenetic movements are required for normal ridge formation (A, left). A first wave of lateral morphogenetic movements results in the compaction of the AER precursor cells onto the ventral surface of the early limb bud and early ventral ectodermal thickening. A second wave compresses this domain to the distal 1/3 of the ventral limb, and a final wave constricts the cells into the densely packed AER. In En1 mutants (A, right), the first morphogenetic waves occur normally. The final wave, however, is markedly inhibited and a secondary compaction process is initiated at the ventroproximal border of the widened mutant AER (A, bottom right). We further suggest that this final wave of ectodermal movements resembles the closing of a zipper (B, ventral view of limb). This process initiates posteriorly and proceeds anteriorly, bringing the ventral domain of the wild-type AER into close proximity to the dorsal domain, which is anchored at the D-V interface (top). In En1 mutant limbs (middle and bottom), the anatomically ventral AER cells take on dorsal characteristics, causing them to become partially anchored and to recruit cells into an ectopic proximoventral rim. The two rims of the bifurcated mutant AER ultimately converge, although in some cases the ventral rim also begins to zip up on itself (middle). If this self-zipping begins at late stages of limb development, the ectopic secondary AER remains contiguous with the primary AER and intersects the latter at a 90° angle (middle, far right); if it begins earlier, the ectopic secondary AER becomes discontinuous with the primary AER and goes on to promote the outgrowth of distinct ventral digits (bottom).

Fig. 6.

Schematic drawing illustrating a model for the morphogenetic movements important in mouse AER development. Based on our results, prior histochemical studies in mouse and lineage data in chick (Milaire,1974; Michaud et al., 1997; Altabef et al., 1997), we propose that cells which give rise to the mouse AER initially overlie much of the presumptive limb mesoderm and that sequential, convergent morphogenetic movements are required for normal ridge formation (A, left). A first wave of lateral morphogenetic movements results in the compaction of the AER precursor cells onto the ventral surface of the early limb bud and early ventral ectodermal thickening. A second wave compresses this domain to the distal 1/3 of the ventral limb, and a final wave constricts the cells into the densely packed AER. In En1 mutants (A, right), the first morphogenetic waves occur normally. The final wave, however, is markedly inhibited and a secondary compaction process is initiated at the ventroproximal border of the widened mutant AER (A, bottom right). We further suggest that this final wave of ectodermal movements resembles the closing of a zipper (B, ventral view of limb). This process initiates posteriorly and proceeds anteriorly, bringing the ventral domain of the wild-type AER into close proximity to the dorsal domain, which is anchored at the D-V interface (top). In En1 mutant limbs (middle and bottom), the anatomically ventral AER cells take on dorsal characteristics, causing them to become partially anchored and to recruit cells into an ectopic proximoventral rim. The two rims of the bifurcated mutant AER ultimately converge, although in some cases the ventral rim also begins to zip up on itself (middle). If this self-zipping begins at late stages of limb development, the ectopic secondary AER remains contiguous with the primary AER and intersects the latter at a 90° angle (middle, far right); if it begins earlier, the ectopic secondary AER becomes discontinuous with the primary AER and goes on to promote the outgrowth of distinct ventral digits (bottom).

Wnt7a and En1 are required only for late regulation of Lmx1b

D-V patterning in the early limb involves complex interactions between mesoderm and ectoderm. Tissue rotation experiments in chick suggested that D-V polarity of the early limb field is initially specified by mesoderm and, just prior to limb outgrowth, control is transferred to the overlying ectoderm (Geduspan and MacCabe, 1989). Given these results, we were somewhat surprised to find that expression of the dorsalizing gene Lmx1b is not restricted to early dorsal limb mesenchyme, but expression extends into the ventral mesenchyme. However, by stage 2, Lmx1b expression has receded from the ventral mesenchyme, resulting in distinct dorsal and ventral molecular identities of the limb mesoderm.

Analysis of En1 and Wnt7a mutants demonstrated that late, but not early, Lmx1b regulation in mouse is dependent on En1 and Wnt7a. In pre-AER En1 mutant limbs, the lack of En1 expression and persistent Wnt7a expression ventrally has little apparent effect on initiation of Lmx1b expression and its subsequent dorsal restriction. However, ventral reexpression of Lmx1b does occur in En1 mutants, at stages 5-6, and this is dependent on Wnt7a expression as it does not occur in En1/Wnt7a double mutants. Furthermore, in Wnt7a mutant limbs, Lmx1b expression is normal until stages 5-6, when it is lost dorsally, but only in distal regions, consistent with a study in chick demonstrating selective distal loss of Lmx1b expression after removal of the Wnt7a expressing dorsal ectoderm (Riddle et al., 1995).

Loss of En1 leads to aberrant ventral AER maturation and the formation of ectopic AERs

Our mutant studies have demonstrated that En1 is not required for the process of AER induction or the initial stages of AER formation. Loss of En1 function has no apparent effect on the early thickening of the pre-AER ventral ectoderm (stage 0.5-1)or on the initial induction of AER marker genes, such as Dlx2 and Fgf8. Furthermore, at stage 3, the cells of the dorsal AER of En1 mutant limbs, like those of wild-type limbs, do not express Wnt7a and they stratify and initiate Fgf4 expression, an AER differentiation marker (Niswander et al., 1994; Chan et al., 1995; Haramis et al., 1995).

In contrast, differentiation of the ventral portion of the AER, where En1 is normally expressed, appears to be delayed and abnormal in En1 mutants. At the time when a mature AER is apparent in wild-type limbs, the ventral portion of the En1 mutant AER remains extended with the anterior region being much broader than the posterior. This difference in the En1 mutant AER phenotype along the A-P axis likely reflects an exaggeration of the inherent A-P asymmetry characteristic of normal ridge constriction, which begins posteriorly and proceeds anteriorly (see Fig. 2G; Milaire, 1974). The ventral AER cells of stage 3/4 En1 mutants remain relatively flat and, while they express Dlx2 and Fgf8, they do not express Fgf4 (Fig. 2L) and do not repress expression of Wnt7a until late (Loomis et al., 1996 and Fig. 2B,F). Nevertheless, the ventral AERs of En1 mutants appear to have some characteristics of a functional AER, since the underlying mesenchyme proliferates. In fact, the expanded En1 mutant ventral AER at these stages appears to secrete more growth factors than normal, since the anterior ventrodistal mesenchyme is visibly wider than normal and it expresses genes usually restricted to AER-associated distal mesenchyme, such as Msx2 (data not shown).

Roughly coincident with the up-regulation of the dorsalizing gene Lmx1b in the ventral mesenchyme of stage 5/6 En1 mutant limbs, the cells along the ventral border of the expanded AER begin to thicken (Fig. 6A) and occasionally initiate Fgf4 expression. Morphologically, the AER of En1 mutants appears bifurcated with the two rims being markedly further apart anteriorly than posteriorly. In some cases (∼30%), independent ectopic AERs form, which display prolonged expression of AER marker genes and go on to promote either independent digit outgrowths or partial distal digit duplications on the ventral surface of En1 mutants limbs.

Based on these observations, we propose a model that likens the final phase of AER constriction to the closing of a zipper (Fig. 6B). In wild-type limb buds, the two halves of the zipper are the dorsal and ventral AER domains. The zipping occurs between stages 1 and 3 with the dorsal half remaining relatively fixed and the ventral half being pulled towards it in a posterior-to-anterior direction. In En1 mutants, differentiation of the ventral half of the AER is markedly delayed and the cells eventually acquire characteristics of dorsal AER cells. Although a zipping process occurs in the normal posterior-to-anterior direction in En1 mutants, the cells in the middle of the broadened AER are pulled toward both the dorsal and ventral AER borders due to the transformation of the ventral half of the AER into a functionally ‘dorsal’ AER. In some cases, distinct secondary AERs form, at right angles to the normal AER axis, due to self-zipping of the ectopic rim (Figs 2B, 4D-H, 6B).

Role of Wnt7a in ectopic AER formation in En1 mutants

The finding that, in En1/Wnt7a double mutants, bifurcated AERs do not form, nor do later ectopic AERs or ventral digits, demonstrates that Wnt7a is required for initiation of abnormal secondary AER formation in En1 mutants. In addition, the finding supports our suggestion that the early abnormal AER structures seen in En1 mutants later produce ectopic ventral digits. Since the AER of En1/Wnt7a double mutants is only slightly broader than normal, this indicates that the early ventral expression of Wnt7a in En1 mutants plays a role in the delay in compaction of the ventral AER cells. This could be due to early expression of Wnt7a in the ventral AER; for example, if Wnt7a prevents necessary cell-shape changes. Alternatively, if during normal limb development Wnt7a secreted from the dorsal non-AER ectoderm is involved in specifying the adjacent AER cells to take on a dorsal phenotype then, in En1 mutants, the ectopic signalling of Wnt7a ventrally would lead to specification of the ventral AER cells as dorsal. Since one apparent characteristic of dorsal AER cells is to stay anchored and pull or recruit cells toward them, this could account for the initial delay in compaction of the En1 mutant ventral AER cells and the later formation of a bifurcated AER.

With respect to the question of the genetic mechanism by which Wnt7a expression leads to formation of secondary AERs in En1 mutants, it is interesting to note that the position of the secondary AER thickening roughly coincides with both the timing of, and the contour of, the ectopic ventral Lmx1b expression domain. Furthermore, in the En1 mutants, ectopic ventral digits and AERs never develop in the posterior autopod where ventral Lmx1b expression is not maintained (Fig. 4J). Since in En1/Wnt7a double mutants Lmx1b is not reexpressed in the ventral mesenchyme, it is possible that Lmx1b also plays a role in the stratification of the ventroproximal AER border and formation of a bifurcated AER.

In Wnt7a mutants, the dorsal AER is only slightly broader than normal, indicating that Wnt7a alone does not normally play a major role in AER formation. Since, Lmx1b is expressed in the dorsal mesenchyme of Wnt7a mutant limbs at the time of AER formation, it is possible that Lmx1b might compensate for the lack of Wnt7a expression in the dorsal ectoderm. Alternatively, or in addition, another Wnt gene might partially compensate for loss of Wnt7a dorsally. It will be interesting to determine whether the Notch pathway also regulates the later process of AER maturation and/or earlier process of AER induction.

In summary, En1 is essential for the normal maturation of the ventral AER, as well as ventral limb patterning. The proper compaction and stratification to form a mature AER requires En1. Our studies indicate that the process of forming ectopic AERs in En1 mutants does not involve a recapitulation of the entire process of AER formation, but instead involves duplication of part of the later process of AER differentiation. We suggest that this formation of a second dorsal AER domain and focal ectodermal thickening requires the juxtaposition of presumptive AER ectoderm and functionally dorsal non-AER ectoderm or mesoderm. These studies have provided a unique example of a genetic pathway impacting on both cell fate and morphogenetic movements, which combine to generate form and pattern.

We thank E. Harris for excellent technical help; Sheila Bell and Bill Scott for provocative discussions; R. Johnson, G. Martin, C. Mahon and M. Morasso for sharing plasmids; Andy McMahon and J. Cygan for sharing mice, plasmids and advice; D-L Song and C. Breigel for help developing a PCR assay for En1 mutants; and C. Breigel for comments and review of the manuscript. A. L. J. is a Howard Hughes Investigator and C. A. L. has been supported by a Dermatology Foundation Career Development Award and an NIH Training Grant number ARO7190T32A7190.

Akita
,
K.
(
1996
).
The effect of the ectoderm on the dorsoventral pattern of epidermis, muscles and joints in the developing chick leg: a new model
.
Anatomy & Embryology
193
,
377
386
.
Altabef
,
M.
,
Clarke
,
J. D. W.
and
Tickle
,
C.
(
1997
).
Dorso-ventral ectodermal compartments and origin of apical ectodermal ridge in developing chick limb
.
Development
124
,
4547
4556
.
Bulfone
,
A.
,
Kim
,
H. J.
,
Puelles
,
L.
,
Porteus
,
M. H.
,
Grippo
,
J. F.
and
Rubenstein
,
J. L.
(
1993
).
The mouse Dlx-2 (Tes-1) gene is expressed in spatially restricted domains of the forebrain, face and limbs in midgestation mouse embryos
.
Mech. Dev.
40
,
129
140
.
Chan
,
D. C.
,
Laufer
,
E.
,
Tabin
,
C.
and
Leder
,
P.
(
1995
).
Polydactylous limbs in Strong’s Luxoid mice result from ectopic polarizing activity
.
Development
121
,
1971
1978
.
Crossley
,
P. H.
and
Martin
,
G. R.
(
1995
).
The mouse Fgf8 gene encodes a family of polypeptides and is expressed in regions that direct outgrowth and patterning in the developing embryo
.
Development
121
,
439
451
.
Crossley
,
P. H.
,
Minowada
,
G.
,
MacArthur
,
C. A.
and
Martin
,
G. R.
(
1996
).
Roles for FGF8 in the induction, initiation, and maintenance of chick limb development
.
Cell
84
,
127
136
.
Cygan
,
J. A.
,
Johnson
,
R. L.
and
McMahon
,
A. P.
(
1997
).
Novel regulatory interactions revealed by studies of murine limb pattern in Wnt-7a and En-1 mutants
.
Development
124
,
5021
5032
.
Davis
,
C. A.
,
Holmyard
,
D. P.
,
Millen
,
K. J.
and
Joyner
,
A. L.
(
1991
).
Examining pattern formation in mouse, chicken and frog embryos with an En-specific antiserum
.
Development
111
,
287
298
.
Gardner
,
C. A.
and
Barald
,
K. F.
(
1992
).
Expression patterns of engrailed-like proteins in the chick embryo
.
Dev. Dynamics
193
,
370
388
.
Geduspan
,
J. S.
and
MacCabe
,
J. A.
(
1987
).
The ectodermal control of mesodermal patterns of differentiation in the developing chick wing
.
Dev. Biol.
124
,
398
408
.
Geduspan
,
J. S.
and
MacCabe
,
J. A.
(
1989
).
Transfer of dorsoventral information from mesoderm to ectoderm at the onset of limb development
.
Anatomical Record
224
,
79
87
.
Grieshammer
,
U.
,
Minowada
,
G.
,
Pisenti
,
J. M.
,
Abbott
,
U. K.
and
Martin
,
G. R.
(
1996
).
The chick limbless mutation causes abnormalities in limb bud dorsal-ventral patterning: implications for the mechanism of apical ridge formation
.
Development
122
,
3851
3861
.
Hanks
,
M.
,
Wurst
,
W.
,
Anson-Cartwright
,
L.
,
Auerbach
,
A. B.
and
Joyner
,
A. L.
(
1995
).
Rescue of the En-1 mutant phenotype by replacement of En-1 with En-2 [see comments]
.
Science
269
,
679
682
.
Haramis
,
A. G.
,
Brown
,
J. M.
and
Zeller
,
R.
(
1995
).
The limb deformity mutation disrupts the SHH/FGF-4 feedback loop and regulation of 5′ HoxD genes during limb pattern formation
.
Development
121
,
4237
4245
.
Johnson
,
R. L.
,
Riddle
,
R. D.
and
Tabin
,
C. J.
(
1994
).
Mechanisms of limb patterning
.
Curr. Opin. Genet. Dev.
4
,
535
542
.
Jurand
,
A.
(
1965
).
Ultrastructural aspects of early development of the fore-limb buds in the chick and the mouse
.
Proc. Royal Soc.
162
,
387
405
.
Kelley
,
R. O.
and
Fallon
,
J. F.
(
1983
).
A freeze-fracture and morphometric analysis of gap junctions of limb bud cells: initial studies on a possible mechanism for morphogenetic signalling during development
.
Progress in Clinical & Biological Research
110
,
119
130
.
Laufer
,
E.
,
Dahn
,
R.
,
Orozco
,
O. E.
,
Yeo
,
C. Y.
,
Pisenti
,
J.
,
Henrique
,
D.
,
Abbott
,
U. K.
,
Fallon
,
J. F.
and
Tabin
,
C.
(
1997
).
Expression of Radical fringe in limb-bud ectoderm regulates apical ectodermal ridge formation [see comments]
.
Nature
386
,
366
373
.
Laufer
,
E.
,
Nelson
,
C. E.
,
Johnson
,
R. L.
,
Morgan
,
B. A.
and
Tabin
,
C.
(
1994
).
Sonic hedgehog and Fgf-4 act through a signaling cascade and feedback loop to integrate growth and patterning of the developing limb bud
.
Cell
79
,
993
1003
.
Lindsell
,
C. E.
,
Shawber
,
C. J.
,
Boulter
,
J.
and
Weinmaster
,
G.
(
1995
).
Jagged: a mammalian ligand that activates Notch1
.
Cell
80
,
909
917
.
Logan
,
C.
,
Hornbruch
,
A.
,
Campbell
,
I.
and
Lumsden
,
A.
(
1997
).
The role of Engrailed in establishing the dorsoventral axis of the chick limb
.
Development
124
,
2317
2324
.
Loomis
,
C. A.
,
Harris
,
E.
,
Michaud
,
J.
,
Wurst
,
W.
,
Hanks
,
M.
and
Joyner
,
A. L.
(
1996
).
The mouse Engrailed-1 gene and ventral limb patterning
.
Nature
382
,
360
363
.
Lyons
,
K. M.
,
Pelton
,
R. W.
and
Hogan
,
B. L.
(
1990
).
Organogenesis and pattern formation in the mouse: RNA distribution patterns suggest a role for bone morphogenetic protein-2A (BMP-2A
).
Development
109
,
833
844
.
MacCabe
,
J. A.
,
Errick
,
J.
and
Saunders
,
J. W.
, Jr.
(
1974
).
Ectodermal control of the dorsoventral axis in the leg bud of the chick embryo
.
Dev. Biol.
39
,
69
82
.
Mahmood
,
R.
, et al. 
. (
1995
).
A role for FGF-8 in the initiation and maintenance of vertebrate limb bud outgrowth
.
Current Biology
5
,
797
806
.
Martin
,
G. R.
(
1995
).
Why thumbs are up
.
Nature
374
,
410
411
.
Matise
,
M. P.
and
Joyner
,
A. L.
(
1997
).
Expression patterns of developmental control genes in normal and Engrailed-1 mutant mouse spinal cord reveal early diversity in developing interneurons
.
J. Neurosci.
17
,
7805
7816
.
Meinhardt
,
H.
(
1983
).
Cell determination boundaries as organizing regions for secondary embryonic fields
.
Dev. Biol.
96
,
375
385
.
Michaud
,
J. L.
,
Lapointe
,
F.
and
Le Douarin
,
N. M.
(
1997
).
The dorsoventral polarity of the presumptive limb is determined by signals produced by the somites and by the lateral somatopleure
.
Development
124
,
1453
1463
.
Milaire
,
J.
(
1974
). Histochemical aspects of organogenesis in vertrebrates. Part I. The skeletal system. Limb morphogenesis. The sense organs. In
Handbuch der Histochemie
pp.
38
60
.,
Stuttgart
:
Gustav Fischer Verlag
.
Morasso
,
M. I.
,
Mahon
,
K. A.
and
Sargent
,
T. D.
(
1995
).
A Xenopus distal-less gene in transgenic mice: conserved regulation in distal limb epidermis and other sites of epithelial-mesenchymal interaction
.
Proc. Natl. Acad. Sci. USA
92
,
3968
3972
.
Niswander
,
L.
(
1996
).
Growth factor interactions in limb development
.
Ann. N.Y. Acad. Sci.
785
,
23
26
.
Niswander
,
L.
,
Jeffrey
,
S.
,
Martin
,
G. R.
and
Tickle
,
C.
(
1994
).
A positive feedback loop coordinates growth and patterning in the vertebrate limb [see comments]
.
Nature
371
,
609
612
.
Niswander
,
L.
and
Martin
,
G. R.
(
1992
).
Fgf-4 expression during gastrulation, myogenesis, limb and tooth development in the mouse
.
Development
114
,
755
768
.
Niswander
,
L.
and
Martin
,
G. R.
(
1993
).
FGF-4 regulates expression of Evx-1 in the developing mouse limb
.
Development
119
,
287
294
.
Niswander
,
L.
,
Tickle
,
C.
,
Vogel
,
A.
,
Booth
,
I.
and
Martin
,
G. R.
(
1993
).
FGF-4 replaces the apical ectodermal ridge and directs outgrowth and patterning of the limb
.
Cell
75
,
579
587
.
Niswander
,
L.
,
Tickle
,
C.
,
Vogel
,
A.
and
Martin
,
G.
(
1994
).
Function of FGF-4 in limb development. [Review]
.
Mol. Reprod. Dev.
39
,
83
88
.
Noramly
,
S.
,
Pisenti
,
J.
,
Abbott
,
U.
and
Morgan
,
B.
(
1996
).
Gene expression in the limbless mutant: polarized gene expression in the absence of Shh and an AER
.
Dev. Biol.
179
,
339
346
.
Ohuchi
,
H.
,
Shibusawa
,
M.
,
Nakagawa
,
T.
,
Ohata
,
T.
,
Yoshioka
,
H.
,
Hirai
,
Y.
,
Nohno
,
T.
,
Noji
,
S.
and
Kondo
,
N.
(
1997
).
A chick wingless mutation causes abnormality in maintenance of Fgf8 expression in the wing apical ridge, resulting in loss of the dorsoventral boundary
.
Mech. Dev.
62
,
3
13
.
Ohuchi
,
H.
,
Yoshioka
,
H.
,
Tanaka
,
A.
,
Kawakami
,
Y.
,
Nohno
,
T.
and
Noji
,
S.
(
1994
).
Involvement of androgen-induced growth factor (FGF-8) gene in mouse embryogenesis and morphogenesis
.
Biochem. Biophys. Res. Comm.
204
,
882
888
.
Parr
,
B. A.
and
McMahon
,
A. P.
(
1995
).
Dorsalizing signal Wnt-7a required for normal polarity of D-V and A-P axes of mouse limb
.
Nature
374
,
350
353
.
Parr
,
B. A.
,
Shea
,
M. J.
,
Vassileva
,
G.
and
McMahon
,
A. P.
(
1993
).
Mouse Wnt genes exhibit discrete domains of expression in the early embryonic CNS and limb buds
.
Development
119
,
247
261
.
Pautau
,
M. P.
(
1977
).
Vertebrate Limb and Somite Morphogenesis (ed
.
Ede
,
D. A.
,
Hinchliffe
,
J. R.
and
Balls
,
M.
)
Cambridge
:
Cambridge University Press
.
Pautou
,
M. P.
and
Kieny
,
M.
(
1973
).
Interaction ec-mesodermique dans l’establissement de la polarite dorso-ventrale du pied de l’embryon de poulet
.
C.r. hebd Seanc. Acad. Sci., Paris
D277
.
Prahlad
,
K. V.
,
Skala
,
G.
,
Jones
,
D. G.
and
Briles
,
W. E.
(
1979
).
Limbless: a new genetic mutant in the chick
.
J. Exp. Zool.
209
,
427
434
.
Riddle
,
R. D.
,
Ensini
,
M.
,
Nelson
,
C.
,
Tsuchida
,
T.
,
Jessell
,
T. M.
and
Tabin
,
C.
(
1995
).
Induction of the LIM homeobox gene Lmx1 by WNT7a establishes dorsoventral pattern in the vertebrate limb
.
Cell
83
,
631
640
.
Riddle
,
R. D.
,
Johnson
,
R. L.
,
Laufer
,
E.
and
Tabin
,
C.
(
1993
).
Sonic hedgehog mediates the polarizing activity of the ZPA
.
Cell
75
,
1401
1416
.
Robinson
,
G. W.
and
Mahon
,
K. A.
(
1994
).
Differential and overlapping expression domains of Dlx-2 and Dlx-3 suggest distinct roles for Distal-less homeobox genes in craniofacial development
.
Mech. Dev.
48
,
199
215
.
Rodriguez-Esteban
,
C.
,
Schwabe
,
J. W.
,
De La Pena,
J.
,
Foys
,
B.
,
Eshelman
,
B.
and
Belmonte
,
J. C.
(
1997
).
Radical fringe positions the apical ectodermal ridge at the dorsoventral boundary of the vertebrate limb [see comments]
.
Nature
386
,
360
366
.
Ros
,
M. A.
,
Lopez-Martinez
,
A.
,
Simandl
,
B. K.
,
Rodriguez
,
C.
,
Izpisua Belmonte
,
J. C.
,
Dahn
,
R.
and
Fallon
,
J. F.
(
1996
).
The limb field mesoderm determines initial limb bud anteroposterior asymmetry and budding independent of sonic hedgehog or apical ectodermal gene expressions
.
Development
122
,
2319
2330
.
Saunders
,
J. W.
, Jr.
(
1948
).
The proximo-distal sequence of the origin of the parts of the chick wing and the role of th ectoderm
.
J. Exp. Zool.
108
,
363
404
.
Saunders
,
J. W.
, Jr.
and
Gasseling,
M.T.
(
1968
).
Epithelial-Mesenchymal Interactions
(ed.
Fleischmajer
,
R.
and
Billingham
,
R. E.
)
Baltimore
:
Williams and Wilkins
.
Sidow
,
A.
, et al. 
. (
1997
).
Serrate2 is disrupted in the mouse limb-development mutant syndactylism
.
Nature
389
,
722
725
.
Tanaka
,
M.
,
Tamura
,
K.
,
Noji
,
S.
,
Nohno
,
T.
and
Ide
,
H.
(
1997
).
Induction of additional limb at the dorsal-ventral boundary of a chick embryo
.
Dev. Biol.
182
,
191
203
.
Tickle
,
C.
(
1995
).
Vertebrate limb development
.
Curr. Opin. Genet. Dev.
5
,
478
484
.
Vogel
,
A.
,
Rodriguez
,
C.
and
Izpisua-Belmonte
,
J. C.
(
1996
).
Involvement of FGF-8 in initiation, outgrowth and patterning of the vertebrate limb
.
Development
122
,
1737
1750
.
Vogel
,
A.
,
Rodriguez
,
C.
,
Warnken
,
W.
and
Izpisua Belmonte
,
J. C.
(
1995
).
Dorsal cell fate specified by chick Lmx1 during vertebrate limb development
.
Nature
378
,
716
720
.
Wanek
,
N.
,
Muneoka
,
K.
,
Holler-Dinsmore
,
G.
,
Burton
,
R.
and
Bryant
,
S. V.
(
1989
).
A staging system for mouse limb development
.
J. Exp. Zool.
249
,
41
49
.
Wurst
,
W.
,
Auerbach
,
A. B.
and
Joyner
,
A. L.
(
1994
).
Multiple developmental defects in Engrailed-1 mutant mice: an early mid-hindbrain deletion and patterning defects in forelimbs and sternum
.
Development
120
,
2065
2075
.
Yang
,
Y.
and
Niswander
,
L.
(
1995
).
Interaction between the signaling molecules WNT7a and SHH during vertebrate limb development: dorsal signals regulate anteroposterior patterning
.
Cell
80
,
939
947
.
Zeller
,
R.
and
Duboule
,
D.
(
1997
).
Dorso-ventral limb polarity and origin of the ridge: on the fringe of independence?
BioEssays
19
,
541
546
.